I have done that :-) The postman didn't bring the saw-filters today :-(  /Niels, oz9ny
|
good idea!
Dr.?William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? Owner - Operator Big Signal Ranch ¨C K9ZC Staunton, Illinois ? Owner ¨C Operator Villa Grand Piton - J68HZ Soufriere, St. Lucia W.I. Rent it: www.VillaGrandPiton.com
email:??bill@... ?
toggle quoted message
Show quoted text
On Jan 2, 2020, at 3:09 AM, erik@... wrote:
?Looks very good. Why not add the pe4302 attenuator chip? Can be found on ebay for less than 1$/pcs -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Looks very good. Why not add the pe4302 attenuator chip? Can be found on ebay for less than 1$/pcs -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
I am working on a shield for an Arduino Due. Hopefully I can mill and test the first version of the shield today. /Niels, oz9ny 
|
Should be able to put it all down on a single pc board, including sockets for Si4432 modules and a processor. As per the end of post 595. I'll probably build something manhatten style once my modules arrive. Chuck Adams in the old qrp-tech group recommended a sheet metal punch and CEM type PCB material: ? ?? /g/qrp-tech/message/5839? ?? /g/qrp-tech/message/13389The CEM material is much easier on the punch than FR4. Google "Hand Held Power Hole Punch Set Sheet Metal? Die Aviation Heavy Duty" for the punch. Another option for prototyping is to use a spot weld drill bit such as this: ? ?? /g/BITX20/topic/68158902Available in 3/8" and 5/16" sizes. Can be used to cut copper islands into a PC board, or press harder and get a manhatten pad. Rectangular pieces of PC board are often used to create boxes for shielding. Though I wonder if small hunks of thin walled copper pipe might be easier. Ideally, socket the shields into the board so they can be popped off to work on it. Or if low enough, just take off a top cover of EMI gasket material. If all the compartments are the same height, could be just one big cover. But I don't have any experience with what kind of shielding works on something like this. My cellphone works, and that's full duplex into a tower 40 air miles away. So if you know what you are doing, it's apparently possible to? have something well shielded but small. On the other hand, in building his spectrum analyzer, W7ZOI laid out each section to fit into a Hhammond 1590b die-cast box:? ?? The thing about SMA connectors everywhere is you can easily swap stuff in like extra gain just to see what happens.? But if this thing works well and is easily duplicated for $20 in parts, somebody should?make a PCB. Jerry
On Wed, Jan 1, 2020 at 05:56 PM, wb6ogd wrote:
Erik is still doing great engineering and education on this project.. THANK YOU AGAIN ERIK!
I have been slowly receiving parts for mine.
Has anyone else been thinking about how to construct theirs?
I am thinking copper clad sort of egg crate, separate boxes for each stage.? I want to eliminate all the interstage
SMA hardware and make it physically much smaller.
Bad idea?? ANY other thoughts?
Happy New Year!
73,
Gary
WB6OGD
|
Erik is still doing great engineering and education on this project.. THANK YOU AGAIN ERIK!
I have been slowly receiving parts for mine.
Has anyone else been thinking about how to construct theirs?
I am thinking copper clad sort of egg crate, separate boxes for each stage.? I want to eliminate all the interstage
SMA hardware and make it physically much smaller.
Bad idea?? ANY other thoughts?
Happy New Year!
73,
Gary
WB6OGD
|
Re: TinySA and phase noise. Understanding the limitations of the SI4432
#spectrum_analyzer
On Wed, Jan 1, 2020 at 10:19 AM, Jerry Gaffke wrote:
If the RSP1 has a 10 bit ADC, how can it have more than a 6dB * 10 = 60dB dynamic range?
This is done with "decimation" every 4 time reduction in samples/second give 10dB more dynamic range. So even though the ADC is 10MSamples/Second the data is decimated to a lower amount of samples/second to get more dynamic range And yes, you may be right the actual accuracy is 12 bit. The two spurs are indeed generated inside the RSP1 The AGC is always switched off. I tested if the tinySA works better with AGC on but its terrible. The only masking that happens is due to phase noise.? -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: TinySA and phase noise. Understanding the limitations of the SI4432
#spectrum_analyzer
Interesting.? And even educational.
Thanks.
A few likely stupid questions.
Just to let you know I'm paying attention:
If the RSP1 has a 10 bit ADC, how can it have more than a 6dB * 10 = 60dB dynamic range?
The spurs at 9.42 and 10.58 mhz of fig 1 and fig 2 are about the same even though
the signal source is completely different.? So must be from inside the RSP1.
Perhaps aliases due to the sample rate of the of the RSP1's ADC?
Not present in fig 3, using the tinySA.
Fig 3 shows a 2mhz wide spectrum centered on the 10 mhz SA9851 signal
as captured by the Si4432 in the tinySA.? ?Noise floor is nice and flat.
I'm thinking the displayed noise floor should drop when below 9.5 mhz and above 10.5 mhz,
since the 10mhz signal would no longer be getting through the 434 mhz SAW filter.
With no strong carrier present, the AGC in the Si4432 would kick up the gain
ahead of the ADC, so the ADC could work best on the smaller signals.
How am I thinking wrong?
Jerry, KE7ER
toggle quoted message
Show quoted text
On Wed, Jan 1, 2020 at 06:06 AM, <erik@...> wrote:
First I used a RSP1 and an AD9851 to establish the ability of the RSP1 to measure close in phase noise. Apart from some spurs this measurement showed some noise starting (the bump at the bottom of the peak marked with "A" below -70dB below the signal .
|
Any oscillator will not only output one frequency but will show small variations around the chosen frequency. This is called close in phase noise. To understand the limitations of the tinySA I wanted to measure the phase noise of the SI4432. First I used a RSP1 and an AD9851 to establish the ability of the RSP1 to measure close in phase noise. Apart from some spurs this measurement showed some noise starting (the bump at the bottom of the peak marked with "A" below -70dB below the signal .  Then I switched the tinySA to signal generator mode which combines the out of the two SI4432 modules at 434MHz and 444MHz into the mixer to generate a 10MHz output signal. As can be seen in the second picture the hump at the bottom of the peak market with A has increased with 10dB and also became wider. This is caused by the fluctuations of the oscillators in the two SI4432 and what is call close in phase noise.  The practical impact of this phase noise can clearly be seen in the third picture taken from the tinySA measurement of the same SA9851 signal. At the bottom of the peak at 10MHz there is a widening visible. The difference between the second and third picture is caused by the difference in resolution bandwidth.? The RSP1 was measuring with a 500Hz FFT bin size where the tinySA measured with a 30kHz resolution filter bandwidth. To better understand the practical impact I added a signal at -54dBm at an offset around 60kHz. (10.058MHz). As can be seen this close by signal is almost masked by the phase noise of the -20dBm signal at 10MHz. Luckily the SI4432 dynamic range is still very good as there is no sign of overloading. Even when both signals fall within the bandwidth of the 434MHz IF filter of the tinySA the performance of the SI4432 is still good.  -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Re: Comparing the performance of SDRPlay RSP1 with the tinySA for HF measurements
Nice comparison.? It would be interesting to see these tests done with an RSP1A since the RSP1 is out of production (1st generation product).? There were quite a few hardware changes and improvements in the RSP1A and so it has much better performance than the RSP1.
Roger
|
The 2GH SA I build uses the cheap AD88307/AD603 combination for up to 120dB range at 10.7MHz and a bandwidth of 300kHz It proves to be difficult to build a RF/IF chain without AGC and a bandwidth of 300kHz that has a better? dynamic range. -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
There are parts for Wideband RF detection too.? Ill search a couple manufacturers for parts and find something that has a good dynamic range that is also cheap.? There should be some good recent candidates¡ ? ? Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Jerry Gaffke via Groups.Io
Sent: Monday, December 30, 2019 12:39 PM
To: [email protected]
Subject: Re: [HBTE] Can you use a single SI4432 as spectrum analyzer? #spectrum_analyzer ? Dr Schmidt,
Exactly how would you implement this external A/D?
First you'll need a device that detects the RF and creates a signal for your A/D to look at.
Can you suggest an RF detector that can do better than 120 dB of dynamic range?
That's what this $2 Si4432 module might possibly be able to achieve.?
Two of the 120khz B3790 SAW filters will give a low enough stop band that we might be able
to test that 120dB figure.
Though when attempting build something with a dynamic range like that,
I'm sure Eric's correct about learning a whole bunch in the process.
Jerry, KE7ER
On Mon, Dec 30, 2019 at 09:02 AM, K9HZ wrote: Fact of the matter is, implementing an external A/D with more resolution is pretty easy too if needed¡ ?
|
Don't have much experience with Virtualbox, installed it a few times on some client windows boxes.
Most of my servers have at least a few KVM VMs running, newer had any major issues.
Hi,
After I posted my last it occurred to me that it is in the kernel and
when I looked in the right place it is listed there:) I will give it a
try and see if it works any better than VirtualBox. Your comments about
that are welcome, too.
Happy New Year and...
73,
Bill? KU8H
|
Hi,
After I posted my last it occurred to me that it is in the kernel and when I looked in the right place it is listed there:) I will give it a try and see if it works any better than VirtualBox. Your comments about that are welcome, too.
Happy New Year and...
73,
Bill KU8H
toggle quoted message
Show quoted text
On 12/30/19 2:33 PM, Dragan Milivojevic wrote: KVM is in the kernel, you just have to install the frontend tools and libraries from the repositories.
Package install of virt-manager will usually trigger the rest of the dependencies.
On Mon, 30 Dec 2019 at 18:49, Bill Cromwell <wrcromwell@... <mailto:wrcromwell@...>> wrote:
Hi,
You made me look. Maybe not all. I have a lightweight Linux
distribution
on my computer and KVM is not present. I do see some KVM pieces listed
on the repositories. I am getting satisfaction with WINE and
VirtualBox.
If/when I go to a new distribution I will look into this some more.
Note
that I am not running SA.exe nor the associated hardware so this is not
a statement that SA.exe runs on my computer.
Happy New Year and...
73,
Bill? KU8H
--
bark less - wag more
|
KVM is in the kernel, you just have to install the frontend tools and libraries from the repositories.
Package install of virt-manager will usually trigger the rest of the dependencies.?
Hi,
You made me look. Maybe not all. I have a lightweight Linux distribution
on my computer and KVM is not present. I do see some KVM pieces listed
on the repositories. I am getting satisfaction with WINE and VirtualBox.
If/when I go to a new distribution I will look into this some more. Note
that I am not running SA.exe nor the associated hardware so this is not
a statement that SA.exe runs on my computer.
Happy New Year and...
73,
Bill? KU8H
|
Dr Schmidt,
Exactly how would you implement this external A/D?
First you'll need a device that detects the RF and creates a signal for your A/D to look at.
Can you suggest an RF detector that can do better than 120 dB of dynamic range?
That's what this $2 Si4432 module might possibly be able to achieve.?
Two of the 120khz B3790 SAW filters will give a low enough stop band that we might be able
to test that 120dB figure.
Though when attempting build something with a dynamic range like that,
I'm sure Eric's correct about learning a whole bunch in the process.
Jerry, KE7ER
toggle quoted message
Show quoted text
On Mon, Dec 30, 2019 at 09:02 AM, K9HZ wrote:
Fact of the matter is, implementing an external A/D with more resolution is pretty easy too if needed¡
?
|
If you want to go beyond the SI4432 why not?build something like this:  I found there is a huge difference between a concept and something that actually works. Making something work is a tremendous learning experience. And the best way to ensure you really understand what you are doing is to blog about it. -- HBTE Files section:? /g/HBTE/filesErik, PD0EK
|
Hi,
You made me look. Maybe not all. I have a lightweight Linux distribution on my computer and KVM is not present. I do see some KVM pieces listed on the repositories. I am getting satisfaction with WINE and VirtualBox. If/when I go to a new distribution I will look into this some more. Note that I am not running SA.exe nor the associated hardware so this is not a statement that SA.exe runs on my computer.
Happy New Year and...
73,
Bill KU8H
toggle quoted message
Show quoted text
On 12/30/19 11:43 AM, Dragan Milivojevic wrote: There is no need for Virtualbox under Linux, KVM is already bundled with all major distributions.
Remapping of virtual ports is a one click operation.
On Mon, 30 Dec 2019 at 16:33, PA3CRZ <j.lammerts@... <mailto:j.lammerts@...>> wrote:
I'm afraid this will never run in Wine/Mono.
Mono is the open source version of .NET, but apparently not
completely so....
The closest thing I could find about this error is this :
If the error I encounter is related to what is being said there,
SA.exe cannot run in Mono.
This is a sad thing, because a lot of people are using Linux
nowadays....
Maybe the alternative can be to install Virtualbox in Linux, and
install windows in a VM.
Just need to figure out how a virtual com port can be mapped onto a
physical one on the host.
73,
Hans
--
bark less - wag more
|
There is no need for Virtualbox under Linux, KVM is already bundled with all major distributions. Remapping of virtual ports is a one click operation.
I'm afraid this will never run in Wine/Mono.
Mono is the open source version of .NET, but apparently not completely so....
The closest thing I could find about this error is this :
If the error I encounter is related to what is being said there, SA.exe cannot run in Mono.
This is a sad thing, because a lot of people are using Linux nowadays....
Maybe the alternative can be to install Virtualbox in Linux, and install windows in a VM.
Just need to figure out how a virtual com port can be mapped onto a physical one on the host.
73,
Hans
|
Fact of the matter is, implementing an external A/D with more resolution is pretty easy too if needed¡ ? ? Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ ? ?
toggle quoted message
Show quoted text
From: [email protected] [mailto:[email protected]] On Behalf Of Jerry Gaffke via Groups.Io
Sent: Monday, December 30, 2019 10:52 AM
To: [email protected]
Subject: Re: [HBTE] Can you use a single SI4432 as spectrum analyzer? #spectrum_analyzer ? I've got a fair bit to learn about the Si4432, but here's my take thus far:
If the Si4432 can be coaxed into giving us a dynamic range of 120dB in detecting a signal,
that's quite good.? I don't see anything yet in the datasheet describing how much of that 120dB
is in the ADC's and how much is had by adjusting the gain of the PGA, best guess is around 50-50.
So when scanning at the maximum bandwidth allowed by the SAW filter and the Si4432 is
not trying to discriminate, we might get 120dB.? Which is great!.
But when using DSP algorithms in the Si4432 to bring the resolution down to 3khz, it would be
limited by the bits in the ADC's, and thus have less dynamic range for whatever happens to be
in the 1mhz or so of spectrum that the SAW filter allows through.? Still a very useful feature.
If by using two of the B3790 SAW filters I can get over 100db of dynamic range
on a 120khz RBW, I'll be very happy.
The datasheet of the B3550 used by Eric can be found here? ??
That guy is now obsolete, Mouser is recommending the B3760 as a replacement.
The B3730 is cheaper than the B3760, and looks to me about as good.
Both of those replacements seem to have a better stopband at 437.5mhz as shown in the graphs.
> RSP1 is a I/Q mixer into a 10bit 10MHz ADC
And the RSP2 has 12 bits, the RSP1A has 14 bits.
But the tinySA can be much cheaper than any of them.
And is fully hackable, which is my primary interest.
Jerry, KE7ER
On Sun, Dec 29, 2019 at 01:29 PM, <erik@...> wrote: My SW ignores the maximum rbw. Thus no problem.
90dB is considerably better than what the Si4432 can do. which I believe is more like 60dB.
Please read the datasheet. The range is120dB
?
|
K9HZ
Roger Need